Crystal structures of the ATP-binding and ADP-release dwells of the V1 rotary motor

V1-ATPases are highly conserved ATP-driven rotary molecular motors found in various membrane systems. We recently reported the crystal structures for the Enterococcus hirae A3B3DF (V1) complex, corresponding to the catalytic dwell state waiting for ATP hydrolysis. Here we present the crystal structures for two other dwell states obtained by soaking nucleotide-free V1 crystals in ADP. In the presence of 20 μM ADP, two ADP molecules bind to two of three binding sites and cooperatively induce conformational changes of the third site to an ATP-binding mode, corresponding to the ATP-binding dwell. In the presence of 2 mM ADP, all nucleotide-binding sites are occupied by ADP to induce conformational changes corresponding to the ADP-release dwell. Based on these and previous findings, we propose a V1-ATPase rotational mechanism model.

dark), and closer (CR; darker) conformations of Eh-A and Eh-B are shown. (m-p) Structures of nucleotide-bound V 1 (bV 1 ), viewed and coloured as in a-d. The bound AMP-PNP molecules are shown in space-filling representation and coloured red.

Supplementary Figure 2
Electron-density maps of the nucleotide-binding sites in the 2 AMP-PNP-bound V 1 complex (2 ATP V 1 ). (a,b) Top views of 2 ATP V 1 from the cytoplasmic side at the level of the nucleotide-binding sites. (c-h) Nucleotide-binding sites of A O (c,d), A C (e,f), and A CR (g,h) of 2 ATP V 1 viewed as in the left panel of Figure 4a. Residues involved in nucleotide binding are shown in a stick representation. Colours correspond to those described in Figure 4. Left panels (a,c,e,g): the 2|Fo|-|Fc| maps contoured at 1.0 sigma are shown in grey. Right panels (b,d,f,h): the |Fo|-|Fc| maps calculated without AMP-PNP:Mg 2+ at the binding pockets contoured at 4.0 sigma are shown in red (negative) and green (positive).

Supplementary Figure 3
Electron-density maps of the nucleotide-binding sites in the 2 ADP-bound V 1 complex (2 ADP V 1 ). (a,b) Top views of 2 ADP V 1 from the cytoplasmic side at the level of the nucleotide-binding sites. (c-h) Nucleotide-binding sites of A O' (c,d), A C (e,f), and A C (g,h) of 2 ADP V 1 viewed as in the left panel of Figure 4a. Residues involved in nucleotide binding are shown in a stick representation. Colours correspond to those described in Figure 4. Left panels (a,c,e,g): the 2|Fo|-|Fc| maps contoured at 1.0 sigma are shown in grey. Right panels (b,d,f,h): the |Fo|-|Fc| maps calculated without ADP:Mg 2+ at the binding pockets contoured at 4.0 sigma are shown in red (negative) and green (positive).

Supplementary Figure 4
Crystal packing contacts of eV 1 , 2 ADP V 1 and 3 ADP V 1 . (a,d,g) Top views from the cytoplasmic sides. (b,e,h) Side views. (c,f,i) Bottom views from the periplasmic side. The structures of eV 1 (a-c), 2 ADP V 1 (d-f), and 3 ADP V 1 (g-i) are shown in surface representation. The residues involved in the crystal packing with buried surface area >10 Å 2 , as calculated by PDBePISA (http://pdbe.org/pisa/), are shown in yellow.

Supplementary Figure 5
Omit electron-density maps of the DF complex and the C-terminal domains of Eh-A and Eh-B of 2 ADP V 1 and 3 ADP V 1 . The structures of 2 ADP V 1 and 3 ADP V 1 that showed conformational changes were validated for possible model bias by making omit maps in the shifted regions, such as the DF complex and the C-terminal domains of the A 3 B 3 complex. Eh-A (a,c,e,g,i,k) or Eh-B (b,d,f,h,j,l) with the DF complex of 2 ADP V 1 (a-f) and 3 ADP V 1 (g-l) viewed as in Figure 3c-h. Colours correspond to those described in Figure 3. The |Fo|-|Fc| maps calculated without the DF complex and the C-terminal domains of the A 3 B 3 complex contoured at 2.0 sigma are shown in grey.  Figure 4a, 6a, 6b, and 7c, respectively.

Supplementary Figure 8
Electron-density maps of nucleotide-binding sites in the 3 ADP-bound V 1 complex (3 ADP V 1 ). (a,b) Top views of 3 ADP V 1 from the cytoplasmic side at the level of the nucleotide-binding sites. (c-h) Nucleotide-binding sites of A HC (c,d), A C (e,f), and A CR' (g,h) of 3 ADP V 1 viewed as in the left panel of Figure 4a. Residues involved in nucleotide binding are shown in a stick representation. Colours correspond to those described in Figure 4. Left panels (a,c,e,g): the 2|Fo|-|Fc| maps contoured at 1.0 sigma are shown in grey. Right panels (b,d,f,h): the |Fo|-|Fc| maps calculated without ADP:Mg 2+ and SO 4 2at the binding pockets contoured at 4.0 sigma are shown in red (negative) and green (positive).

Supplementary Figure 9
Electron-density maps of nucleotide-binding sites in the P i -bound V 1 complexes. (a,d,g,j; left panels), V 1 complex soaked with 20 µM P i (0 Pi V 1 :20 µM). (b,e,h,k; middle panels), V 1 complex soaked with 200 µM P i (0 Pi V 1 :200 µM). (c,f,i,l; right panels), V 1 complex soaked with 2 mM P i (1 Pi V 1 ). (a-c) Top views from the cytoplasmic side at the level of the nucleotide-binding sites. (d-l) Nucleotide-binding sites of A O (d-f), A C (g-i), and A CR (j-l) of these V 1 complexes viewed as described in the left panel of Figure 4a. The bound molecules (P i :Mg 2+ ) and residues involved in nucleotide binding are shown in a stick representation. Colours correspond to those described in Figure 4. The |Fo|-|Fc| maps calculated without P i :Mg 2+ at the binding pockets contoured at 4.0 sigma are shown in red (negative) and green (positive).

Supplementary Figure 12
Comparison of the conformations of tryptophan residues around nucleotide binding sites of 2 ATP V 1 , 2 ADP V 1 , and 3 ADP V 1 . The viewing position and colours of these nucleotide-binding sites correspond to the left panel of Figure 4a. Trp248 (light blue), other tryptophan residues (yellow), AMP-PNP and ADP are shown as stick representations. (a-c) 2 ATP V 1 (colour) is superimposed onto the corresponding subunit (grey) of eV 1 using all atoms. (d-f) 2 ADP V 1 (colour) is superimposed onto the corresponding subunit (grey) of eV 1 using all atoms. (g-i) 3 ADP V 1 (colour) is superimposed onto the corresponding subunit (grey) of eV 1 using all atoms.

Supplementary Figure 14
A molecular mechanism model for 360° rotation of the V 1 complex starting from the catalytic dwell state. The structure models are based on the crystal structures of 2 ATP V 1 (a,d,g), 2 ADP V 1 (b,e,h), and 3 ADP V 1 (c,f,i) determined in this study. C-terminal domain views from the cytoplasmic side are shown. ATP with a yellow 'P' represents an ATP molecule that is committed to hydrolysis, as in Figure 9. ATP binds to the 'bindable-like' form at 0° (b). It is then hydrolysed to ADP and P i at 240° (g). The P i molecule with lower affinity than ADP is released at the same angle, and the conformation changes from 'tight' to 'ADP-bound' (h). The rest of the ADP molecule is also released at the same angle (240°) induced by ATP binding (i). ATPase activity of the purified V 1 -ATPase in the presence of AMP-PNP or ADP was measured by the colourimetric method using molybdic acid. The reaction was initiated by the addition of 1 mM ATP, after a 10 min pre-incubation with various concentrations of AMP-PNP or ADP, and terminated by the addition of 10% sodium dodecyl sulphate. The initial rate of the ATPase reaction at 23 °C was determined within 4 min. ATPase activities of the purified V 1 -ATPase in the presence of various concentrations of sodium phosphate (P i ) were measured using an ATP regenerating system. ATP hydrolysis rates at 23 °C were determined in terms of the rate of NADH oxidation, which was measured as a decrease in absorbance of 340 nm. All data represent means ± standard estimated errors (SEM) of three independent experiments. The 100% value of the ratio corresponds to the averaged ATPase activity of EhV 1 in the absence of the inhibitors.